Bug 105781

Summary: GPE flooding prevention - gpe06 interrupt storm on Macbook Pro (late 2013)
Product: ACPI Reporter: Gökçen Eraslan (gokcen.eraslan)
Component: BIOSAssignee: Lv Zheng (lv.zheng)
Status: CLOSED DUPLICATE    
Severity: normal CC: ben.r.xiao, cengizc, fernando, rui.zhang
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: 4.2.2 Subsystem:
Regression: No Bisected commit-id:
Attachments: acpidump
DSDT
Disassambled DSDT
cat /proc/acpi/wakeup
Full "dmidecode --no-sysfs" output

Description Gökçen Eraslan 2015-10-11 10:50:39 UTC 
Created attachment 189951 [details]
acpidump

I have been experiencing that a kworker thread is hogging CPU for long time:

$ cat /sys/firmware/acpi/interrupts/gpe06
 3058798   enabled

Disabling it solves the problem temporarily:

# echo disable > /sys/firmware/acpi/interrupts/gpe06

There seems to be other people complaining about the same thing:

https://bugzilla.kernel.org/show_bug.cgi?id=85881#c48
https://bugzilla.redhat.com/show_bug.cgi?id=1192856
https://bbs.archlinux.org/viewtopic.php?pid=1500325#p1500325
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1403653
http://teslabs.com/articles/ubuntu-15-04-mbp-11-2/#high-cpu-usage-by-kworker
https://gist.github.com/tombee/c1c225388f261fcddd4b#power-management-issues

acpidump attached, dmidecode gives:

/sys/firmware/dmi/tables/DMI: Unexpected end of file
Table is unreachable, sorry.
# dmidecode 3.0
Getting SMBIOS data from sysfs.
SMBIOS 2.4 present.
45 structures occupying 2597 bytes.
Table at 0x7AD14000.

PS: Version of OS X installed is 10.11.
Comment 1 Gökçen Eraslan 2015-10-11 10:54:20 UTC 
Created attachment 189961 [details]
DSDT

DSDT attached: `sudo cat /sys/firmware/acpi/tables/DSDT`
Comment 2 Gökçen Eraslan 2015-10-11 10:56:19 UTC 
Created attachment 189971 [details]
Disassambled DSDT
Comment 3 Zhang Rui 2015-10-12 05:50:13 UTC 
please attach the full acpidump.
please attach the output of "cat /proc/acpi/wakeup".
Comment 4 Gökçen Eraslan 2015-10-12 18:45:57 UTC 
Created attachment 190101 [details]
cat /proc/acpi/wakeup
Comment 5 Gökçen Eraslan 2015-10-12 18:48:16 UTC 
I added "cat /proc/acpi/wakeup" output. acpidump is full already.
Comment 6 Gökçen Eraslan 2015-10-13 16:35:33 UTC 
Created attachment 190181 [details]
Full "dmidecode --no-sysfs" output
Comment 7 Gökçen Eraslan 2015-10-14 17:00:29 UTC 
Do you need any more information? I cannot remove NEEDINFO flag.
Comment 8 Gökçen Eraslan 2016-01-11 14:10:29 UTC 
I can reproduce it even on android-x86. Is anything else needed?
Comment 9 Lv Zheng 2016-05-16 22:24:51 UTC 
Same issue as Bug 117481, let's track this there.

*** This bug has been marked as a duplicate of bug 117481 ***